Hardware and software used

1. Lidar
2. RGB camera
3. Jetson microcontroller
4. ROS - Robot Operating System
5. Rviz - for visualization
6. Gazebo - for simulation

Software stack

(implemented by each memeber of the team of 5)

1. EKF SLAM for initial mapping
2. Mobilenet trained on COCO dataset to detect and classify food items from RGB images
3. A-star algorithm for path planning
4. Shortest path traversal using travelling salesman approach
5. Differential flatness controller for long range and PI controller for short range motion

Server

1. Socket programming using Winsock library to establish new client connections.
2. User data bound to independednt objects for unique identification.
3. Input texts stored if user offline.
4. All communications made non-blocking.

Client

1. User inputs and server communication made non-blocking using futures.
2. Command table approach to handle special commands like help.
3. Error handling and file compression of previous chats using lzw compression.